Organised Crime Group identified and referred to police: £500k–£1m in dental fraud stopped before further losses

Case 5  ·  Collusion Level  ·  UK + International

WHERE WE STARTED

A large UK dental insurer was processing claims that, individually, fell within normal parameters. Viewed together, they formed the signature of an Organised Crime Group — identical invoices submitted under different patient names, from different regions, by people who shared financial and network connections invisible to any single-claim review.

THE CHALLENGE

The insurer had a small fraud team that had only ever investigated fraud on an individual customer basis. They had no prior experience of organised crime operating at scale across their book, and no infrastructure to detect it.

What had already been tried:

Previous fraud controls were designed for opportunistic individual fraud — a claimant slightly overstating a bill, or a provider submitting a duplicate. Nothing in the existing toolkit was designed to detect coordinated multi-party fraud operating across geographical regions and multiple insurance schemes simultaneously.

HOW KIRONTECH FOUND IT

• Data from the dental insurer was imported into HIP and analysed for treatment frequency and invoice patterns.
• Analysis surfaced implausibly high levels of dental treatments and unusual clustering of treatment patterns across several areas of the country.
• HIP identified service replication: identical invoices being submitted by different people in different regions — a hallmark signature of coordinated fraud that no individual-claim review could surface.
• Network analysis confirmed that many of the suspects shared financial and other connections across multiple insurance schemes — validating the OCG pattern.
• Machine learning predicted further frauds of a similar nature based on the identified pattern — a prediction that was subsequently validated when additional frauds of the same type were found.

WHAT CHANGED

Beyond stopping the immediate losses, this case transformed the insurer’s fraud capability. The team that had only ever handled individual claims cases now had a formal OCG investigation on its record, with an ongoing police engagement and a new intelligence-sharing relationship with law enforcement. The ML model that predicted further frauds was validated in real time — and continues to run across incoming claims.